2.4 OUTER STRUCTURE
The problem of analysing and forecasting the outer circulation of tropical cyclones has received much less attention than any other aspect of structure and structure change. Its importance depends entirely upon the vulnerability of the persons and property at risk. A common example is an airport with a single runway. A cross wind of only 20-30 kt causes no damage by itself but may be enough to prevent aircraft from taking off and escaping the much stronger winds to come. To the airport manager, the extent of 30 kt winds around the tropical cyclone, and their anticipated time of arrival at the field, are important forecasts indeed!
No standard definition exists for outer circulation. The most common definition appears to be the extent of gale force (17 ms-1, 34 kt) winds at the surface. This is typically 100-500 km from the centre and can be highly asymmetric. Another measure is the Radius of Outer Closed Isobar (ROCI), which is related to (but not interchangeable with) the radius of gale-force winds. Climatologies of the ROCI exist for the Atlantic and western North Pacific, where it has been observed to vary between 100 and 1000 km (Merrill, 1982). Both measures are poorly related to intensity (Weatherford and Gray, 1988); it is possible, even common, to find small tropical cyclones with extremely high maximum winds and very large ones with weaker maximum winds.
Figure 2.10: Percentage of large tropical cyclones relative to all systems in the period 1961-1969 that passed through 5o Marsden squares during August, September and October in the western North Pacific (after Merrill, 1982).
Figure 2.11: Percentage of large tropical cyclones relative to all systems in the period 1957-1977 that passed through 5o Marsden square during August, September and October in the North Atlantic (after Merrill, 1982).
All else being equal, filling tropical cyclones tend to be larger than intensifying ones (Merrill, 1982). More details can be found in Holland (1987).
Outer circulation is conceptually one of the easier aspects of tropical cyclone structure to predict. The governing dynamics, particularly for tropical cyclones entering the subtropics, are more nearly those of large-scale mid-latitude systems and do not involve the complex and largely unobservable interactions between vortex, convection, and surface processes which dominate the core region. Unfortunately, lack of directed research to date means that a climatology of the radius of outer closed isobar (ROCI) and a few general principles implied by it are the only forecast aids available at present. Hence this Guide can only present general principles and climatological information.
The outer circulation is probably the easiest element to observe directly and the most difficult to infer remotely from satellite imagery, for which recent attempts have met with mixed results. Because ships skirt tropical cyclones, the peripheral circulation is often well observed if it affects shipping lanes. Although these observations often define the regions where there are no gales, this is useful information for analysis. When defining the outer circulation near land, issues of anemometer exposure and land-sea roughness difference need to be carefully considered; the gale-force isotach defined at sea can be very slow to advance inland.
Figure 2.12: Examples of lower-tropospheric environments associated with a small tropical cyclone (left) and a large, or growing tropical cyclone (right).
Low level satellite cloud motions produced at larger national centres also may be available, either directly or incorporated in analyses transmitted from these centres. These may define the outer circulation well in quadrants free of high cloud, but except in relatively symmetric tropical cyclones they should not be considered representative of other quadrants. Studies of mean conditions in convective bands associated with tropical cyclones (Wei and Gray, 1985) indicate that surface winds are higher in major cloud bands than in clear regions at the same radius.
Passive microwave radiometers, such as the SSM/I aboard the US DMSP series, can be used to estimate surface wind speeds in rain-free areas. These observations are based on scattering from the wind-roughened sea surface and have shown skill in monitoring peripheral winds around tropical cyclones (Rappaport, 1991).
The climatology of ROCI for the North Atlantic and western North Pacific are contained in Figs. 2.10, 2.11. Note the frequency of larger tropical cyclones late in the season and at higher latitudes in both basins; the western North Pacific also has a tendency to produce large systems at low latitudes that is not apparent in the North Atlantic. The following general principles are suggested:
1. The initial size of the tropical cyclone is related to the environment in which it forms. Most tropical cyclones form near monsoon troughs and tend to be large initially. Tropical cyclones forming in trade-wind regimes tend to be smaller;
2. Tropical cyclones often contract as they intensify, and tend to expand as they fill;
3. Subsequent changes in size follow the change in the tropical cyclone's surroundings. Those which move out of the monsoon trough into the trades may contract. Those which approach a mid-latitude surface trough or enter a break in the subtropical ridge often expand as they are sandwiched between surface anticyclones (Fig. 2.12);
4. Very large tropical cyclones are most common in the subtropical latitudes late in the season. The typical synoptic situation is for the tropical cyclone to move poleward ahead of a mid-latitude trough with the subtropical ridge to its east and a more baroclinic anticyclone to the west.
Particularly in cases 3 and 4, but also for monsoon troughs in which a westerly wind surge accompanies tropical cyclone formation, the extent of gale force winds is likely to be highly asymmetric. It is helpful to devote as much attention to predicting the movement and pressure changes in surrounding synoptic systems as to the tropical cyclone itself. For instance, a developing anticyclone will often be associated with an increased pressure gradient and stronger winds between itself and the tropical cyclone and lead to increased asymmetry. Fig. 2.13 provides a schematic life cycle for tropical cyclones that form in a trade-wind regime.
No known statistical methods exist for predicting outer circulation changes. The ability of dynamical models, both limited area or global, to predict tropical cyclone size have not been evaluated either, although their results probably depend heavily upon the size of the bogus vortex if one is specified and they should therefore be used with extreme caution. Because the vortex size is believed to be related to the dynamics of tropical cyclone motion, a recent emphasis on motion research may indirectly lead to more attention to outer circulation analysis and prediction as well.
Figure 2.13: Average changes of size and intensity of 12 recurving Atlantic hurricanes. The circle is the radius of the average outer closed isobar. (Merrill, 1984).
Recommendation: It is recommended that the radius of outer closed isobar, and/or the radius of gale force winds in four quadrants along, and normal to the direction of motion be added to the "best track" archived for tropical cyclones. Within a decade or less, this will enable climatology-persistence class multiple regression methods to be prepared for each tropical cyclone basin. Synoptic conditions associated with very large or very small tropical cyclones, or rapid increases in outer circulation, should also be summarised and catalogued.
Contents Chapter 2.5